There is growing concern within the social work profession that the quantity and quality of current research in social work is inadequate to inform the major practice areas and meaningfully enhance the existing knowledge base. Various perspectives on this problem were recently presented in the report of the Task Force on Social Work Research (National Institute of Mental Health, 1991). The report concluded that social work research is currently in a state of crisis and made recommendations for curriculum changes, additional funding sources, and new staffing for key national social work organizations. The gap between social work practice and research was identified as one of the major barriers to the development of practice-relevant research. A range of different approaches to integrating these two specialty areas were suggested and described by participants of another recent social work research conference (Jenkins, 1992). Subsequently, Lindsey and Kirk (1993) questioned whether social work research is truly in crisis or whether it suffers from a chronic condition that reflects a historical underlying ambivalence toward scientific inquiry: "The social work profession's historical and ideological bias . . . favors action over theory, practice insight over scientific research, and good intentions over effective outcomes; this has led to a de-emphasis on research" (p. 378). These authors point specifically to the lack of expansion of social work research into leading private and public research universities such as Duke, Stanford, Harvard, Northwestern, Princeton, Cornell, Vanderbilt, Virginia, or Arizona as illustrative of the problem. They therefore recommend the development of increased research specialization at selected schools of social work that would become major social work research centers. Both reports agree on the seriousness of the problem for the future of social work as a profession, and they recommend consideration of a range of new strategies to bring about fundamental change in the profession in this area. The research effort described here represents one such strategy. RESEARCH IN HOSPITALS Historically two approaches to the development of research in hospitals have been used. Most frequently, in hospitals where social work research is done it is embedded in quality assurance activities (Berkman & Weissman 1983). Another model, reported by Turnbull, Saltz, and Gwyther (1988), by Simon (1991), and by Blythe and Tripodi (1989), emphasizes teaching research to social work clinicians to enhance their ability to understand and use research methodologies in their practices. A third model, based on our experience over a 13-year period, is presented in this article. This model emphasizes the establishment of a research unit as an integral part of a hospital social work department. Each of these models addresses the research issues differently. Of the three models, the one we present requires the most extensive commitment of resources. Its emphasis is on basic research such as developing knowledge of the processes of coping and adaptation and of the interventions most effective in enhancing these processes. The other two models focus more on management goals. In this third model, published research findings not only are a goal, but also define the success or failure of such a program. Our goal was to establish a unit that would ultimately be self-sustaining, funded primarily by federal grants and philanthropy. Another goal, which was partially met, was to develop a research program that would result in a mutually interdependent relationship between clinicians and researchers. To this end the research program was planned to undertake studies that would inform practice and program development to enhance the quality of life of patients and their families. For example, areas that were investigated over time included studies of the psychosocial sequelae of cancer survivorship, a broad range of studies of patients with acquired immune deficiency syndrome (AIDS), studies of preventive interventions in childhood bereavement, evaluation of interventions to reduce the prevalence of unmet concrete needs among outpatients with cancer, investigation of the psychosocial sequelae of being a caregiver to a seriously ill individual, and descriptive studies of the range of needs of cancer patients. …